def predictor_wed_server(payload):
model = payload['model']
num = payload['num']
j = parse_json(j=payload)
x, player_info = j.get_Info(num)
new_client = client()
# load specified model
if model == 'NN':
new_client.load_model(f'Models/NN_Model_{num}')
else:
new_client.load_model(f'Models/SVM_Model_{num}.pkl')
# predict
prediction = new_client.predict(x)
# format prediction
player_info['team'] = 'Blue'
if player_info['team'] == 'CHAOS':
prediction = 1 - prediction
player_info['team'] = 'Red'
prediction = "{:.2f}".format(prediction * 100)
prediction = f'{player_info["summoner_name"]}, {player_info["champion"]} has a {prediction}% of winning on the {player_info["team"]} team after {num} minutes since the game\'s start.'
return prediction
def NN_Model(num):
# get info depending on 10 or 15
path, label, d = get_info(num)
# preprocess csv file
newClient = client(path)
newClient.preprocess(label=label, d=d)
newClient.train_NN()
return newClient.model
from Code.client import client
# dense_layers = [3,4]
# layer_sizes = [32, 64, 128]
# drop_ratios = [0.6, 0.8]
dense_layers = [3]
layer_sizes = [32]
drop_ratios = [0.4]
# preprocess data
train_path = "/Users/rostamvakhshoori/Desktop/MatchTimelinesFirst10.csv"
newClient = client(train_path)
newClient.preprocess(
label='blueWins',
d=['gameId', 'blueWins', 'blueTotalExperience', 'redTotalExperience'])
newClient.train_NN()
newClient.analysis()
# train_path = "/Users/rostamvakhshoori/Desktop/MatchTimelinesFirst15.csv"
# X_train, X_test, y_train, y_test = preprocess(train_path)
# logisitic_regression_model = LogisticRegression()
# logisitic_regression_model.fit(X_train, y_train)
# print(logisitic_regression_model.score(X_test, y_test))
# clf = svm.SVC()
# clf.fit(X_train, y_train)
# print(clf.score(X_test, y_test))
# for dense_layer in dense_layers:
# for layer_size in layer_sizes: